MusicGPT icon

MusicGPT

Interact with MusicGPT API for audio processing and voice management

Actions25

Audio Generation → Denoise

AI-generated

Summary

Submits an audio file to the MusicGPT API's denoise endpoint to remove background noise. It starts an asynchronous job and returns task identifiers, not the cleaned audio.

Inputs

  • Audio Source — Selects where the audio comes from: 'URL' (default) or 'Binary Property' (a file from a previous node).
  • Audio URL (required) — URL of the audio file to process. Shown and required when Audio Source is URL.
  • Binary Property Name (required) — Name of the input item's binary property that holds the audio file (default 'data'). Shown and required when Audio Source is Binary Property. The file is uploaded as multipart form data, using its original filename or 'audio.mp3' if it has none.
  • Webhook URL — Optional callback URL that MusicGPT notifies when processing completes. It is sent only if non-empty.

Output shape

One JSON item per input item, with the fields success, task_id, conversion_id and eta from the MusicGPT API response.

The call only queues the job, so the response has no denoised audio. To get the result, use the Webhook URL, or poll with Search > Get Conversion by ID (type 'Denoising') using the returned task_id or conversion_id. With 'Continue on Fail' enabled, a failed item returns an item with an 'error' message instead. Each output item is paired to its input item.

Examples

Example 1: Clean up a noisy recording hosted at a public URL.

Audio Source = URL; Audio URL = https://example.com/interview.mp3; leave Webhook URL empty and poll for the result by conversion_id.

Example 2: Denoise a file downloaded by an earlier HTTP Request node and get notified when done.

Audio Source = Binary Property; Binary Property Name = data; Webhook URL = https://your-webhook.com/callback.

Links

Discussion